[ic] question on mod_interchange socket closing (different from previous question)

[Kevin Walsh]

Jeff Dafoe [jeff@badtz-maru.com] wrote:
> 
> I have been looking at the mod_interchange source code in order to
> examine it for a potential performance advantage it may give to IC that I
> theorize it might offer.  If I couple my examination of ic_transfer_response
> with some behaviors I have noticed with my system, I start to think that in
> cases where ic_transfer_response returns before getting to the bottom of
> the subroutine and IC doesn't close the socket itself, the IC socket may not
> be closed for a really long time.  Since this is only a theory based on a
> weak understanding of the Apache API, and it represents an insignificant
> issue, just an offhand "it's possible" or "it's probably not likely" is all
> I am looking for.
> 
It seems that you have found a valid concern in mod_interchange.
I've made a change that should fix the problem.  Please take a look
at the new mod_interchange-1.25, which you'll find on my RTFM site.
I'll run a few more tests on the code before committing to the
Interchange 4.9 CVS archive.

I'm reasonably sure that Apache would have cleaned up the open socket
for us, but I agree that it's best that the module does it anyway.

Thanks for reporting your findings.  I'll be interested to hear any
other comments you have.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/